The family Cyatheaceae (s. str.) defined here is a group with scaly stems. The
classification of this scaly group has a long history, but views on the generic rank are contro-
versial. The division of three genera based mainly on indusia had lasted for almost one cen-
tury and was finally proved unnatural. Not all species lacking indusia are closely related to
each other. In other words, among the exindusiate species, some are closer to those of indusiate
species than to the other exindusiate ones, and vice versa. Obviously, the loss of indusia must
have taken place in different evolutionary lines. In the last few decades, Tryon (1970, 1982)
made a new chart on the basis of the structure of petiole scales, in which six scaly genera were
proposed, while Holttum (1957, 1963, 1974, 1981), with the emphasis on the chromosome number
n=69 in all species so far counted, insisted on one genus in a broad seuse, including two su-
bgenera and two sections in each subgenus.
After Chinese material has been studied, the author proposes several points as follows: 1)
It is acceptable that the family is divided into two main groups based on the structure of
petiole scales. New information from Chinese material supports this separation. Pneumathodes,
a kind of aerating tissue along lateral sides of raches and stipes, are of two forms. If they come
to the surface, the surface is broken and a row of fusiform holes emerge at the maturity, when
we call them open pneumathodes. If they do not come to the surface, no hole but a discontinous
whitish line apears on the surface, when we call them closed pneumathodes. These two
forms of pneumathodes are correlated with the two patterns of petiole scales. 2) Sporangia of
two kinds, producing 16 or 64 spores, occur in the family. For Chinese species, one group
is of 16-spored sporangia and is considered as a derivative, and the others are of 64-spored
sporangia as most ferns. 3) Among various forms of indusia, two evolutionary changes may
be traced. although the parallel development of indusia in different groups indicates that the
structure is not a useful systematic character for distinguishing genera. In one line the primi-
tive hemitelioid indusium fully developed finally forming a spherical one covering a whole
sorus, and in the other it simplified forming a scalelike indusium hidden by a mature sorus.
Perhaps, the loss of indusia results from simplifying.
Recognized in the present paper are 14 species and 2 varieties, which are grouped into
three taxa: Sphaeropteris, Alsophila and Gymnosphaera. Sphaeropteris, characterized by closed
pneumathodes and structurally conform petiole scales, is given the generic status. Alsophila
and Gymnosphaera are both with open pneumathodes and structurally marginate petiole scales,
but the former is with 16-spored sporangia, while the latter with 64-spored sporangia. They
are so closely related in morphology that they hardly merit generic rank, as compared with the
genus Sphaeropteris which has 64 spores in each sporangium. They are thus treated as two sub-
genera under the genus Alsophila.
Cyatheaceae has a pantropical distribution extending northwards to S. China and S. W.
China. Its northern limit in China is at the Huaying Mountain in Linshui County of Sichuan
Province, at about 30°N.
I am indebted very much to Prof. R. G. Ching and K. H. Shing for stimulating discussions
of the classification of Cyatheaceae, and also to Prof. W. M. Chu and H. S. Kung for very
helpful comments on the manuscript.
